|Publication number||US7945497 B2|
|Application number||US 11/961,380|
|Publication date||May 17, 2011|
|Filing date||Dec 20, 2007|
|Priority date||Dec 22, 2006|
|Also published as||US20080154651, US20110218827, WO2008079325A1|
|Publication number||11961380, 961380, US 7945497 B2, US 7945497B2, US-B2-7945497, US7945497 B2, US7945497B2|
|Inventors||Timothy P. Kenefick, Eric Besman, Theresa C. Murphy|
|Original Assignee||Hartford Fire Insurance Company|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (119), Non-Patent Citations (53), Referenced by (20), Classifications (7), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application claims the benefit of U.S. Provisional Application Ser. No. 60/876,684, filed Dec. 22, 2006, the entire contents of which are incorporated herein by reference.
In general, the invention relates to the field of computerized predictive models and their use for intelligent decision making. More specifically, the invention relates to predictive models in which outputs of predictive models serve as inputs to one another to form a predictive feedback loop and methods for using such models.
Effectively identifying appropriate insurance coverages for a particular customer requires skill and experience developed over years of practice. Those having such skill are limited in number and thus provide only limited bandwidth to insurers. These limitations make it difficult for a customer to obtain coverage guidance in a real-time fashion on demand. Thus, a need exists in the art for a computerized system capable of identifying appropriate coverages for customers with reduced need for expert human interaction.
Moreover, many requests for insurance are relatively routine. Expending underwriter energy on evaluating such requests presents a burden on insurance companies with limited underwriting bandwidth and slows the process for customers of obtaining insurance and insurance quotes. Thus, a need exists in the art for a system to identify insurance requests that can be handled with no or reduced human underwriter involvement.
Accordingly, in one aspect, the invention relates to a system for handling insurance requests, such as new coverage applications, renewal applications, and insurance quote requests. The system includes a user interface for accepting data related to an insurance request from a user. The system also includes at least two computerized predictive models whose outputs each serve as inputs to the other. A business logic module is included in the system to process the accepted data by the predictive models to determine one or more sets of recommended policy parameters for the user and to output the determination to the user via the user interface. In various embodiments, the predictive models also output workflow components (such as underwriting workflow components) and price components (such as premium prices) corresponding to output suggested sets of coverage parameters.
In one embodiment, the system includes a data capture module that extracts input parameters from a variety of data sources for the predictive models. Suitable data sources include third party data sources, providers of telematics data, and unstructured text, which can be mined using a text mining engine and/or an information extraction engine.
In another embodiment, the business logic module of the system includes a rule engine configured to identify parameters for computerized predictive models for which additional data would improve their performance. The identified parameters are then output to the data capture module and/or the user interface.
According to another aspect, the invention relates to a method for handling an insurance request. The method includes accepting data related to an insurance request from a user via a user interface. A business logic module then processes the accepted data to determine a recommended set of policy parameters for an insured by executing first and second predictive models based on the accepted data. At least one output of the first predictive model serves as an input parameter to the second computerized predictive model and at least one output of the second computerized predictive model serves as an input to the first computerized model. The method also includes outputting, by the business logic module, the recommended set of policy parameters to the user via the user interface.
According to a third aspect, the invention relates to a computer readable medium storing computer readable instructions, which, when executed by a processor, cause the processor to carry out the method and system functionality described above.
The foregoing discussion will be understood more readily from the following detailed description of the invention with reference to the following figures.
To provide an overall understanding of the invention, certain illustrative embodiments will now be described. However, it will be understood by one of ordinary skill in the art that the systems and methods described herein may be adapted and modified as is appropriate for the application being addressed and that the systems and methods described herein may be employed in other suitable applications, and that such other additions and modifications will not depart from the scope hereof.
System 100 includes an insurance carrier system 102 in communication with user systems 104 and 106 and data sources 108 and 110. User system 104 is an agent system operated by an insurance agent. User system 106 is a customer system operated by an insurance customer. Data sources 108 are third party sources of data, including, for example, government entities (such as departments of motor vehicles, federal and state census bureaus, and meteorological services), private data vendors, news sources, and credit bureaus. Using such sources, the insurance system 102 may obtain census information, motor vehicle registration and driver information, weather information, property information (such as construction type, construction date, etc.), crime rates, credit histories, financial information, etc. The data may be aggregated data for geographies, industries, or demographics, or it might be specific to current or prospective customers. For example, the third party data may include, without limitation, information about company size, finances, and industry. Illustrative third party text documents include news stories, product information, analyst reports, and documents related to medical treatments, including devices, procedures, and agents.
Data sources 110 are sources of telematics data, i.e., data derived from sensors monitoring property to be insured. The telematics data sources 110 may include monitoring services engaged to monitor property of an insurance customer, systems operated by the insurance customer to monitor the sensors, and sensors that directly output to the insurance carrier system 102. For example, sensor may monitor the condition and/or use of individual or fleets of vehicles, insured property, and goods, or the condition and/or location of individuals. For example, with respect to automobile insurance, telematics data may include, without limitation, speed, location, acceleration, deceleration, environmental conditions (e.g., presence of icy roads or precipitation), tire pressure, and engine use time. For insured structures, the data may include, without limitation, temperature, humidity, alarm system status, smoke alarm status, and air quality. For individuals, telematics data might include, without limitation, blood pressure and pulse. For insured goods, the data may include, without limitation, the location of the goods, data related to the handling of the goods, and data related to environment surrounding the goods.
The insurance carrier system 102 communicates with the user systems 104 and 106 and the data sources 108 and 110 over one or a combination of communication networks 112. The communication networks 112 may be wired, wireless, or a combination thereof. They may be publicly accessible, such as the Internet, or part of a private communications network. Preferably, communications over the communication networks 112 are encrypted to protect the privacy of insurance customers.
The user systems 104 and 106 may be any computing device known in the art, such as a desktop computer, laptop computer, or personal digital assistant capable of networked communication and having a microprocessor capable of executing the user interfaces described herein. In particular, the user systems 104 and 106 include software for providing a user interface to the respective agent or insurance customer. For the agent system 104, the software may include a thin or thick client, or it may take the form of browser executable code for providing a user interface via a standard Internet browser software application. For the insurance customer system 106, the software preferably takes the form of browser executable code for providing a user interface via a standard Internet browser software application. The user interfaces provided by the user systems 104 and 106 are described further below. In addition to communicating with agent systems 104 and user systems 106, the insurance system 102 may also received insurance requests from insurance aggregators, brokers, an e-agent interface and/or Internet portals intended for identifying insurance coverage for customers. In some embodiments, these alternative providers of an insurance requests are associated with one or more agents.
The insurance carrier system 102 includes a plurality of application servers 114, a plurality of load balancing proxy servers 116, a data warehouse 118, a business logic processor 120, and one or more carrier terminals 122. These computing devices are connected by a local area network 124.
The application servers 114 are responsible for interacting with the user systems 104 and 106. For example, the application servers 114 include software for generating web pages for communication to the user systems 104 and 106. These web pages serve as user interfaces for the insurance agent and insurance customer, respectively, to interact with the insurance carrier system 102. Alternatively, or in addition, one or more of the application servers 114 may be configured to communicate with thin or thick clients operating on the agent system 104. The load balancing proxy servers 116 operate to distribute the load among application servers 114.
The data warehouse 118 is the main electronic depository of the insurance company's current and historical data. The data warehouse 118 includes one or more interrelated databases that store information relevant to insurance data analysis. The interrelated databases store both structured and unstructured data. In one implementation, one or more of the interrelated databases store electronic copies of insurance forms, either in an image format or a text-searchable format keyed to a customer or claim. Other forms of unstructured data stored in the data warehouse 118 include, without limitation, audio files storing insurance customer telephone conversations, transcripts of such conversations, image files, e.g., documenting the condition of insured property, and video files. Other databases in the interrelated databases store data, for example, in a relational database, in various data fields keyed to various identifiers, such as, without limitation, customer, data source, geography, or business identifier (such as Standard Industry Classification Code). The information stored in the data warehouse 118 is obtained through communications with customers, agents, vendors, and third party data providers and investigators.
The business logic processor 120 includes one or more computing devices tasked with carrying out the primary insurance decision analyses carried out by the insurance company. In particular, the business logic processor 120 is configured to make suggestions of policy parameters to potential customers and current customers seeking new coverage or renewal of existing coverage. The business logic processor 120 is further configured to determine a workflow to be followed by the insurance company to evaluate the customer for such coverage, and to determine a corresponding price. These determinations are preferably made using one or more computerized predictive models. The business logic processor 120 is described further in relation to
The insurance carrier terminals 122 provide various user interfaces to insurance carrier employees to interact with the business logic processor 120. The interfaces include, without limitation, interfaces to add new insurance coverages and coverage options to the insurance carrier system 102 and to adjust the predictive models utilized by the business logic processor 120. Such interfaces may be integrated into one or more websites for managing the insurance carrier system 102 presented by the application servers 114, or they may be integrated into thin or thick software clients. The carrier terminals 122 can be any computing devices suitable for carrying out the processes described above, including personal computers, lap top computers, personal digital computers, and other computing devices with general purpose processors.
Computing device 201 comprises at least one central processing unit (CPU) 202, at least one read-only memory (ROM) 203, at least one communication port or hub 204, at least one random access memory (RAM) 205, and one or more databases or data storage devices 206. All of these later elements are in communication with the CPU 202 to facilitate the operation of the computing device 201. The computing device 201 may be configured in many different ways. For example, computing device 201 may be a conventional standalone server computer or alternatively, the function of server may be distributed across multiple computing systems and architectures.
Computing device 201 may be configured in a distributed architecture, wherein databases and processors are housed in separate units or locations. Some such servers perform primary processing functions and contain at a minimum, a general controller or a processor 202, a ROM 203, and a RAM 205. In such an embodiment, each of these servers is attached to a communications hub or port 204 that serves as a primary communication link with other servers 207, client or user computers 208 and other related devices 209. The communications hub or port 204 may have minimal processing capability itself, serving primarily as a communications router. A variety of communications protocols may be part of the system, including but not limited to: Ethernet, SAP, SAS™, ATP, BLUETOOTH™, GSM and TCP/IP.
The CPU 202 comprises a processor, such as one or more conventional microprocessors and one or more supplementary co-processors such as math co-processors. The CPU 202 is in communication with the communication port 204 through which the CPU 202 communicates with other devices such as other servers 207, user terminals 208, or devices 209. The communication port 204 may include multiple communication channels for simultaneous communication with, for example, other processors, servers or client terminals. Devices in communication with each other need not be continually transmitting to each other. On the contrary, such devices need only transmit to each other as necessary, may actually refrain from exchanging data most of the time, and may require several steps to be performed to establish a communication link between the devices.
The CPU 202 is also in communication with the data storage device 206. The data storage device 206 may comprise an appropriate combination of magnetic, optical and/or semiconductor memory, and may include, for example, RAM, ROM, flash drive, an optical disc such as a compact disc and/or a hard disk or drive. The CPU 202 and the data storage device 206 each may be, for example, located entirely within a single computer or other computing device; or connected to each other by a communication medium, such as a USB port, serial port cable, a coaxial cable, a Ethernet type cable, a telephone line, a radio frequency transceiver or other similar wireless or wired medium or combination of the foregoing. For example, the CPU 202 may be connected to the data storage device 206 via the communication port 204.
The data storage device 206 may store, for example, (i) a program (e.g., computer program code and/or a computer program product) adapted to direct the CPU 202 in accordance with the present invention, and particularly in accordance with the processes described in detail hereinafter with regard to the CPU 202; (ii) databases adapted to store information that may be utilized to store information required by the program. Suitable databases include data warehouse 118 of
The program may be stored, for example, in a compressed, an uncompiled and/or an encrypted format, and may include computer program code. The instructions of the program may be read into a main memory of the processor from a computer-readable medium other than the data storage device 206, such as from a ROM 203 or from a RAM 205. While execution of sequences of instructions in the program causes the processor 202 to perform the process steps described herein, hard-wired circuitry may be used in place of, or in combination with, software instructions for implementation of the processes of the present invention. Thus, embodiments of the present invention are not limited to any specific combination of hardware and software.
Suitable computer program code may be provided for performing numerous functions such as responding to insurance requests, including suggesting insurance policy coverage parameters, calculating a corresponding premium price, and determining a corresponding workflow component for processing the insurance request using the suggested parameters. The program also may include program elements such as an operating system, a database management system and “device drivers” that allow the processor to interface with computer peripheral devices (e.g., a video display, a keyboard, a computer mouse, etc.).
The term “computer-readable medium” as used herein refers to any medium that provides or participates in providing instructions to the processor of the computing device (or any other processor of a device described herein) for execution. Such a medium may take many forms, including but not limited to, non-volatile media and volatile media. Non-volatile media include, for example, optical, magnetic, or opto-magnetic disks, such as memory. Volatile media include dynamic random access memory (DRAM), which typically constitutes the main memory. Common forms of computer-readable media include, for example, a floppy disk, a flexible disk, hard disk, magnetic tape, any other magnetic medium, a CD-ROM, DVD, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, a RAM, a PROM, an EPROM or EEPROM (electronically erasable programmable read-only memory), a FLASH-EEPROM, any other memory chip or cartridge, or any other medium from which a computer can read.
Various forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to the processor 202 (or any other processor of a device described herein) for execution. For example, the instructions may initially be borne on a magnetic disk of a remote computer 208. The remote computer 208 can load the instructions into its dynamic memory and send the instructions over an Ethernet connection, cable line, or even telephone line using a modem. A communications device 204 local to a computing device (or, e.g., a server) can receive the data on the respective communications line and place the data on a system bus for the processor. The system bus carries the data to main memory, from which the processor retrieves and executes the instructions. The instructions received by main memory may optionally be stored in memory either before or after execution by the processor. In addition, instructions may be received via a communication port as electrical, electromagnetic or optical signals, which are exemplary forms of wireless communications or data streams that carry various types of information.
As previously discussed with reference to
As used herein, a “module” may be implemented in software for execution by various types of processors. An identified module of executable code may, for instance, comprise one or more physical or logical blocks of computer instructions which may, for instance, be organized as an object, procedure, or function. Nevertheless, the executables of an identified module need not be physically located together, but may comprise disparate instructions stored in different locations which, when joined logically together, comprise the module and achieve the stated purpose for the module.
Indeed, a module of executable code could be a single instruction, or many instructions, and may even be distributed over several different code segments, among different programs, and across several memory devices. Similarly, operational data may be identified and illustrated herein within modules, and may be embodied in any suitable form and organized within any suitable type of data structure. The operational data may be collected as a single data set, or may be distributed over different locations including over different storage devices, and may exist, at least partially, merely as electronic signals on a system or network. In addition, entire modules, or portions thereof, may also be implemented in programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices or the like or as hardwired integrated circuits.
A predictive model 402 preferably takes into account a large number of parameters. The predictive models 402, in various implementation, include one or more of neural networks, Bayesian networks (such as Hidden Markov models), expert systems, decision trees, collections of decision trees, support vector machines, or other systems known in the art for addressing problems with large numbers of variables. Preferably, the predictive models 402 are trained on prior data and outcomes known to the insurance company. The specific data and outcomes analyzed vary depending on the desired functionality of the particular predictive model. The particular data parameters selected for analysis in the training process are determined by using regression analysis and/or other statistical techniques known in the art for identifying relevant variables in multivariable systems. The parameters can be selected from any of the structured data parameters stored in the data warehouse 118, whether the parameters were input into the system originally in a structured format or whether they were extracted from previously unstructured text.
Business logic module 302 includes an underwriting workflow model 402 a, a coverage recommendation model 402 b, a pricing model 402 c, and a fraud detection model 402 d. The predictive models 402 a-402 d result in the output of an estimated policy price 404, a suggested set of policy parameters 406, and a required underwriting workflow component 408.
The underwriting workflow model 402 a determines one or more underwriting workflow components that will be required to extend a given set of coverages at a particular price to a customer. Coverages for more homogenous industries require less underwriting than coverages for businesses that entail greater individual risk variation. Underwriting work flow components may include, without limitation, a loss-control investigation, review by a human underwriter, verification of data with documentary evidence, automatic issuance/declination of a policy, flagging of additional risk characteristics requiring further exposure development, identification of unique risks that warrant special attention (e.g. review by home office staff or manuscript endorsements), and indication of underwriting authority levels. In addition, or in the alternative, the underwriting workflow model 402 a may output that no separate underwriting steps are required and that the proposed coverage may be extended without further review.
The coverage recommendation model 402 b outputs a suggested set of policy parameters in response to the insurance system 102 receiving an application for a new or renewal policy or an insurance quote request. The policy parameters include, without limitation, suggested coverages and riders, and associated limits and deductibles. The set of policy parameters is derived for the customer based on historical coverage selections and claims histories of other customers having similar attributes, similar claims histories, and similar risk profiles as the applicant for the new or renewal policy. In one implementation, the coverage recommendation model 402 b determines a set of suggested policy parameters dynamically each time a new or renewal policy is requested or a quote request is received. In another implementation, on a periodic basis, the coverage recommendation model 402 b is executed to generate a set of suggested policies. Then, when an applicant requests a new or renewal policy or a quote there for, the coverage recommendation model 402 b recommends a policy from the previously generated set of policies that most closely matches the profile of the applicant. In both cases, a user of the system, for example, an insurance agent or the customer, can then change the suggested policy parameters to personalize the policy to their own preferences. In one implementation, a policy having a suggested set of policy parameters is treated preferentially by the underwriting business logic in comparison to manually generated policies or modified suggested policies.
The pricing model 402 c outputs a price associated with a given set of coverage parameters, underwriting workflow components, and customer attributes. The output may be an actual premium associated with the input, for example, if no manual underwriting is required. Alternatively, the output may be an estimated premium that may vary based on the results of any further underwriting workflow components.
The fraud detection model 402 d outputs a value corresponding to the likelihood that data entered into the insurance system 102 is fraudulent. The value may be a likelihood score, a probability, or a qualitative or quantitative classification, e.g., high, medium, or low, top 30%, middle 30%, or bottom 40% of insurance requests ranked based on their respective likelihoods of fraud.
In one implementation, the predictive models 402 a-402 d operate independently from one another, using only the input received from the data capture module 306 and the data warehouse 312. In the illustrative implementation, the outputs of one or more of the predictive models 402 a-402 d serve as inputs to others of the predictive models 402 a-402 d. In a simple case, the output of the fraud detection model 402 d, which includes a likelihood that the requester of insurance is submitting fraudulent information, serves as an input to the underwriting workflow model 402 a. In response to the fraud detection model 402 d outputting an elevated likelihood of fraud, the underwriting workflow model 402 a is more likely to output a more stringent underwriting process. In this case, the fraud detection model 402 d and the underwriting workflow model 402 a, can be executed serially.
In contrast, in a more complex implementation, two or more predictive models may execute in a loop fashion. For example, the output of the coverage recommendation model 402 b serves as an input to the pricing model 402 c, and the output of the pricing model serves as an input to the coverage recommendation model 402 b. The models 402 b and 402 c execute one after another until their outputs across iterations stabilizes, until the variation in the outputs exhibits desirably small changes, or for a predetermined number of iterations.
For example, in addition to price (i.e., the output of pricing model 402 c), the coverage recommendation model 402 b includes the following input parameters obtained by the data capture module 306: customer size, customer annual revenues, customer assets, business type identifier (e.g., an SIC code), and claims history. Based on the type of business the customer participates in and the customer's claims history, the coverage recommendation model 402 b might recommend a first set of coverages if price were not an option. However, as the amount of insurance a customer is capable and willing to pay will vary based on its revenues and the assets at risk, a coverage recommendation can be improved if price is considered along with these and the other parameters described above. Thus, after an initial set of recommended coverage parameters is determined by the coverage recommendation model 402 b, the output set of coverage parameters is processed along with the data capture module 306 provided parameters to determine a price corresponding to extending such coverage to the particular customer. The price output by the pricing model 402 c is then re-input into the coverage recommendation model 402 b. Depending on the price, the coverage recommendation model may vary its output to recommend additional coverages, fewer coverages, different deductibles and liability limits, and/or different exclusions. The subsequent recommended set of coverages is input into the pricing model 402 c to determine an updated price. The process continues for a predetermined number of iterations or until the output of the coverage recommendation model 402 b or the pricing model 402 c stabilizes.
In an alternative implementation, the coverage recommendation model 402 b outputs a plurality of sets of recommended coverage parameters along with corresponding scores. Each of the recommended coverage sets is then processed in a loop fashion with the pricing model 402 c to tune the parameters of each respective set of coverage parameters. In still another embodiment, the processing loop also includes the underwriting workflow model 402 a, to determine the appropriate underwriting workflow for each recommended set of coverages and corresponding price.
Other suitable parameters for processing by the predictive models 402 a-402 c include, without limitation, location, desired premium range, number of employees, property to be insured, value of property to be insured, products, regulatory regime, number of insurance request resubmissions (as described below), agent identifier, agent profitability history, agent data quality history, employee driving records, wage levels of employees, proximity to certain geographic characteristics, business demographics, and fit of specific operations of a business within a more general classification.
The final outputs of the model are output to the user interface 304 for review.
The user interface screen shot 500 also includes fields 510 for each recommended coverage set 502 a for a user to accept a presented parameter. For presented parameters the user does not wish to accept, the user may select an alternative using drop down boxes 511 or may leave the parameter field as a variable. After the user is finished adjusting the parameters, the user may select a re-submit button 512 a-512 c corresponding to the respective recommended coverage set 502 a-502 c. The revisions are then output back to the business logic module 302 through the data capture module 306. The predictive models 402 a-402 d of the business logic module 302 re-execute using the newly submitted data in addition to prior submitted data. Manually revised parameters are considered fixed, or in some cases fuzzy, criteria by the models 402, and are not varied or are only minimally varied during the re-execution process. For example, if the user modifies a price corresponding to a suggested set of parameters, the models 402 consider the modified price a fixed or fuzzy threshold and will iterate until they find one or more sets of coverage parameters and underwriting workflows that has a price that is near, at, or below that threshold. If after a predetermine number of iterations, no such set of coverages and underwriting workflows are identified, the request is denied. Similarly, in addition, or in the alternative, a user may confirm the selection of a particular coverage, deductible, coverage limit, and/or underwriting workflow component prior to reprocessing by the business logic module 302.
After reviewing the recommended coverage parameter sets 502 a-502 c, the user may save one or more of the recommended coverage sets 502 a-502 c by selecting corresponding save buttons 514 a-514 c. In addition, a user may accept a recommended coverage set 502 a-502 c to submit an application for insurance having the identified parameters by selecting a corresponding accept button 516 a-516 c. Upon acceptance, depending on the desired underwriting workflow component 506, various steps may be automatically initiated. For example, upon selection of accept button 516 a, for which no additional underwriting is required, the insurance system binds the customer to a policy with the presented parameters. Upon selection of accept buttons 516 b or 516 c, the insurance request is automatically forwarded to an underwriting system for assigning personnel to and scheduling manual underwriting review and/or loss control investigations. While the illustrative screen shot 500 only includes three suggested parameter sets 502 a-502 c, alternative user interface 304 implementations may present fewer or a larger number of suggested parameter sets 502 to a user without departing from the scope of the invention.
Referring back to
Referring back to
For data to be obtained from the third party data sources 108 and telematics data providers 110, in one implementation, the data capture module 305 includes a web services server 604 executing one or more web services for automatically querying and obtaining query responses from remote servers operated by the third party data sources 108 and telematics data providers 110. In addition, or in the alternative, the web services server 604 includes a web crawling application to identify relevant information from the broader World Wide Web.
For data to be obtained from user interface 304 users, the data capture module includes a web server 606 for generating web pages to serve to agent systems 104 or customer systems 106 and to receive data therefrom. The web pages may be implemented in part in HTML, SHTML, AJAX, XML, JAVA, MACROMEDIA FLASH, or other web programming or scripting language.
In one implementation, the questions displayed in the web page output by the web server 606 to a user via the user interface 304 are determined dynamically at least in part based on direction from the business logic module 302. The business logic module 302 may direct the web server 606 to generate a web having questions dynamically selected based on data reviewed from the data warehouse 118/312, third party data sources 108, telematics data providers 110, and/or based on responses to prior questions. In one particular implementation, the business logic module 302 selects questions to present to the user via the user interface 304 based on information known about the user. For example, if the user is known to be a customer with limited knowledge of insurance, the business logic module 302 directs the web server 606 to generate a web page that presents questions in a first format; whereas when the user is known to be an experienced agent, the questions are presented in a second format. Similarly, the business logic module 302 may select questions to be presented to the user based on prior history of interactions with the user. For example, if the user is an agent with only limited experience selling policies for the company, the business logic module 302 may direct the web server 606 to generate a web page that presents more or different questions than if the user is known to be an experienced agent. Questions may also be selected by the business logic module 302 based on the performance history of policies issued by a particular agent. For example, if polices issued by a given agent historically result in higher than average loss profiles, the business logic module 302 may direct the web server 606 to generate a web page that presents more or different questions to that particular agent.
The data capture module 306 also includes a text mining engine 608, an information extraction engine 610, and a relationship engine 612 for extracting additional data from unstructured data and for making connections between data entered in relation to one insurance request or claim with other insurance requests and claims. The text mining engine 608, information extraction engine 610, and relationship engine 612 may process documents retrieved from the data warehouse 118, documents retrieved from third party data sources, and the free-form text entry fields of insurance forms presented via the user interface 304.
The text mining engine 608 extracts structured data from the unstructured text included in unstructured text documents received or stored by the insurance system 102. The text mining engine 608 includes an optional optical character recognition process to detect text in documents stored in an image format. In one embodiment, the text mining engine 608 includes a non-natural language parser for identifying key words in documents. The key words identified may be based on a predetermined list of words, or they may be identified by analyzing the frequency of the word in the document or a corpus of documents being analyzed. In another implementation, the text mining engine 608 includes a natural language parser for extracting semantic meaning from text in addition to detecting the presence and/or frequency of particular key words. The text mining engine 608 may take on a number of other forms without departing from the scope of the invention.
The information extraction engine 610, in one implementation, is integrated with the text mining engine 608. The information extraction engine 610 identifies names of people, places, and things in documents and identifies semantic relationships between people and objects. The output of the text mining engine 608 and the information extraction engine 610 may be stored in the data warehouse 118/312 or may be output directly to the business logic module 302.
The relationship engine 612 analyzes data stored in the data warehouse 118/312 to draw linkages between individual datum which may not already be logically linked. The linked data may be related to a single customer or claim, or they may be related to different customers or claims. For example, in handling claims for a first customer, the insurance company learns that a particular individual owned a company that experienced numerous claims. The fact that this individual is associated with a negative loss history is stored in the data warehouse. In a later insurance request from a new customer, the insurance company learns that the new business is owned by that same individual without being directly informed that the individual was associated with the negative loss history. By processing the information by the relationship engine 612, the individual will be linked to their loss history, and the data stored for the new customer will be updated in the data warehouse 118 accordingly.
The relationship engine 612 can also be employed to detect discrepancies in data received from multiple sources. For example, if in a form a customer indicates that an insured property is of a first size, and a third party data source 108, for example, a real estate information database, indicates that the property is of a second size, the relationship engine 612 can instruct the data capture module 306 to obtain further information to verify the actual size or to require investigation of the facts during underwriting. Such discrepancies may be output to the fraud detection model 402 d in the business logic module 302. Alternatively, the relationship engine 612 can issue an alert. Similarly, in analyzing automobile accidents, the relationship engine can detect discrepancies between written accounts of the accident from different parties and data collected from telematics data sources 110 monitoring one or more of vehicles involved in the accident.
Preferably, the insurance decision making system is dynamic in nature. That is, based on information learned from analyses carried out by the predictive models 402 and the relationship engine 612, the text mining engine 608, the information extraction engine 610, the relationship engine 612, and the predictive models 402 are updated to reflect relevant information. For example, the predictive models can be used to detect trends in input data. For example, by analyzing extracted text in relation to outcomes, the predictive models can determine new structured parameters to include in an analysis and/or new weights to apply to previously known parameters. In addition, as new actual data is collected, for example, the actual losses associated with a particular policy are experienced, the system can be retrained with the new outcome data to refine its analysis capabilities. In one implementation, the system is retrained on a monthly basis. In other embodiments, the system is trained on a weekly, quarterly, and annual basis. In another embodiment, the system is retrained continuously.
Next, additional data is obtained to support the handling of the insurance request (steps 702-708). Each of the following steps is optional. At step 704, structured data is extracted from unstructured text using a text mining application. This step may also include extracting information from the text using an information extraction application. The unstructured text may be received via the user interface, from databases operated by the insurance company, such as the data warehouse 118, or third party data sources. At step 706, data is collected from a third data source as those described above in relation to third party data sources 108. At step 708, data is collected from providers of telematics data 110.
Next, the method 700 includes identifying parameters for which no, little, suspect, or otherwise inadequate data was provided that would benefit the proper execution of the handling process (step 710). If any such parameters are identified, the data acquisition steps 702-708 are repeated to obtain data for the identified parameters. If no parameters are identified at step 710, the obtained data is processed using predictive models (step 712). Preferably, the output of a first predictive model serves as an input to a second predictive model. In addition, an output of the second predictive model also serves as an input to the first predictive model. During or after processing, the method may optionally include returning to step 710 to identify if obtaining data for any model data parameters would improve performance of the models.
Based on the processing carried out at step 712, several outputs are provided (steps 714-718). At step 714, one or more sets of suggested policy parameters are output for presentation to the user. At 716, corresponding underwriting workflow components are output in relation to each respective set of suggested policy parameters. Finally, at step 718, corresponding premium prices are also output for presentation to the user via the user interface.
The invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The forgoing embodiments are therefore to be considered in all respects illustrative, rather than limiting of the invention.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5613072||Jun 1, 1995||Mar 18, 1997||Risk Data Corporation||System for funding future workers compensation losses|
|US5712984||May 11, 1993||Jan 27, 1998||Risk Data Corporation||System for funding future workers' compensation losses|
|US5796932||Jul 20, 1995||Aug 18, 1998||Strategic Weather Services||User interface for graphically displaying the impact of weather on managerial planning|
|US5926800||Apr 24, 1995||Jul 20, 1999||Minerva, L.P.||System and method for providing a line of credit secured by an assignment of a life insurance policy|
|US5950150||Sep 8, 1997||Sep 7, 1999||Lloyd; Steven J.||Fire/life safety system operation criteria compliance verification system and method|
|US5970464||Sep 10, 1997||Oct 19, 1999||International Business Machines Corporation||Data mining based underwriting profitability analysis|
|US6014632||Apr 15, 1997||Jan 11, 2000||Financial Growth Resources, Inc.||Apparatus and method for determining insurance benefit amounts based on groupings of long-term care patients with common characteristics|
|US6067488||Dec 2, 1996||May 23, 2000||Data Tec Co., Ltd.||Vehicle driving recorder, vehicle travel analyzer and storage medium|
|US6078857||Dec 23, 1997||Jun 20, 2000||Hyundai Motor Company||Apparatus for deciding a shift pattern suitable for a driver's driving habit using neural network operation and fuzzy inference and a control method thereof|
|US6112225||Mar 30, 1998||Aug 29, 2000||International Business Machines Corporation||Task distribution processing system and the method for subscribing computers to perform computing tasks during idle time|
|US6154658||Mar 11, 1999||Nov 28, 2000||Lockheed Martin Corporation||Vehicle information and safety control system|
|US6163770||Aug 25, 1998||Dec 19, 2000||Financial Growth Resources, Inc.||Computer apparatus and method for generating documentation using a computed value for a claims cost affected by at least one concurrent, different insurance policy for the same insured|
|US6204757||May 28, 1999||Mar 20, 2001||Francis Christopher Evans||Seatbelt usage and safety data accounting system|
|US6246934||May 18, 2000||Jun 12, 2001||Toyota Jidosha Kabushiki Kaisha||Vehicular data recording apparatus and method|
|US6438547||May 8, 2000||Aug 20, 2002||Firepond, Inc.||Computer-readable data product for managing sales information|
|US6594579||Aug 6, 2001||Jul 15, 2003||Networkcar||Internet-based method for determining a vehicle's fuel efficiency|
|US6611740||Mar 14, 2001||Aug 26, 2003||Networkcar||Internet-based vehicle-diagnostic system|
|US6643578||Sep 9, 2002||Nov 4, 2003||Alfred B. Levine||Vehicle drive override system|
|US6710738||Nov 4, 2002||Mar 23, 2004||Lunareye, Inc.||Triggerable remote controller|
|US6756915||Oct 25, 2001||Jun 29, 2004||Lg Electronics Inc.||Geographic information service apparatus and method using a DSRC network|
|US6768417||Feb 28, 2002||Jul 27, 2004||Hitachi, Ltd.||On-vehicle emergency report apparatus, emergency communication apparatus and emergency report system|
|US6832141||Oct 25, 2002||Dec 14, 2004||Davis Instruments||Module for monitoring vehicle operation through onboard diagnostic port|
|US6839305||Feb 16, 2001||Jan 4, 2005||Neil Perlman||Habit cessation aide|
|US6853956||Feb 11, 2003||Feb 8, 2005||Smart Start Inc.||Sobriety testing apparatus having OBD-II connection capability|
|US6917952||Nov 20, 2001||Jul 12, 2005||Burning Glass Technologies, Llc||Application-specific method and apparatus for assessing similarity between two data objects|
|US6920379||Dec 30, 2002||Jul 19, 2005||Renesas Technology Corp.||Vehicle-onboard DSRC apparatus|
|US6925425||Oct 12, 2001||Aug 2, 2005||Motorola, Inc.||Method and apparatus for vehicle operator performance assessment and improvement|
|US6957133||May 8, 2003||Oct 18, 2005||Reynolds & Reynolds Holdings, Inc.||Small-scale, integrated vehicle telematics device|
|US6959296||Sep 13, 2001||Oct 25, 2005||Paul Reiss||Systems and methods of choosing multi-component packages using an expert system|
|US6965326||Jul 21, 2003||Nov 15, 2005||General Motors Corporation||Automated electronic module configuration within a vehicle|
|US6973319||Nov 28, 2001||Dec 6, 2005||Nec Corporation||System and method for measuring traffic flow|
|US6977612||Jun 29, 2004||Dec 20, 2005||Electronic Data Systems Corporation||System and method for wireless asset tracking|
|US6985922||Dec 21, 2001||Jan 10, 2006||S.J. Bashen, Inc.||Method, apparatus and system for processing compliance actions over a wide area network|
|US7007245||Mar 12, 2004||Feb 28, 2006||Chevron U.S.A. Inc.||Product selection expert system|
|US7039592||Mar 28, 2001||May 2, 2006||Pamela S. Yegge||Agricultural business system and methods|
|US7072841||Apr 29, 1999||Jul 4, 2006||International Business Machines Corporation||Method for constructing segmentation-based predictive models from data that is particularly well-suited for insurance risk or profitability modeling purposes|
|US7215255||Jan 21, 2004||May 8, 2007||Bernard Grush||Method and apparatus for a satellite positioning-based metering system for use in transport-related applications|
|US7702615 *||Nov 4, 2005||Apr 20, 2010||M-Factor, Inc.||Creation and aggregation of predicted data|
|US20010020229||Jul 30, 1998||Sep 6, 2001||Arnold Lash||Method and apparatus for determining high service utilization patients|
|US20020010598||Dec 18, 2000||Jan 24, 2002||Johnson Jerome Dale||System and method for providing configuration and sales information to assist in the development of insurance plans|
|US20020072958||Oct 24, 2001||Jun 13, 2002||Takuya Yuyama||Residual value forecasting system and method thereof, insurance premium calculation system and method thereof, and computer program product|
|US20020099596||Nov 26, 2001||Jul 25, 2002||Geraghty Michael Kevin||Dynamic ratemaking for insurance|
|US20020147617||Apr 24, 2001||Oct 10, 2002||Michael Schoenbaum||Health cost calculator/flexible spending account calculator|
|US20020161609||Oct 23, 2001||Oct 31, 2002||Zizzamia Frank M.||Commercial insurance scoring system and method|
|US20020173885||Mar 13, 2001||Nov 21, 2002||Lowrey Larkin Hill||Internet-based system for monitoring vehicles|
|US20020194113||Dec 21, 2001||Dec 19, 2002||Abb Group Services Center Ab||System, method and computer program product for risk-minimization and mutual insurance relations in meteorology dependent activities|
|US20020198801||Jan 7, 2002||Dec 26, 2002||Dixon Deborah A.||System and method for asset accumulation and risk management|
|US20030028406||Jul 24, 2002||Feb 6, 2003||Herz Frederick S. M.||Database for pre-screening potentially litigious patients|
|US20030061075||May 17, 2002||Mar 27, 2003||Converium Reinsurance (North America) Inc.||System and method for rating and structuring bands of crop production insurance|
|US20030097330||Nov 18, 2002||May 22, 2003||Amway Corporation||System and method for detecting fraudulent transactions|
|US20030101080||Nov 28, 2001||May 29, 2003||Zizzamia Frank M.||Method and system for determining the importance of individual variables in a statistical model|
|US20030105651||Nov 30, 2001||Jun 5, 2003||Edward Gendelman||Process for insuring and risk managing the decommissioning and/or abandonment of an oil and gas production facility|
|US20030167191 *||Feb 25, 2002||Sep 4, 2003||Slabonik Elizabeth Ann||System and method for underwriting review in an insurance system|
|US20030171956||Mar 6, 2002||Sep 11, 2003||Cox Steven R.||System for improving logistics, tracking and billing for worker's compensation insurance|
|US20040103002||Oct 23, 2003||May 27, 2004||Colley John Lawrence||Method for health plan management|
|US20040128147||Dec 26, 2002||Jul 1, 2004||Sundar Vallinayagam||Method and system to implement complex pricing rules|
|US20040133439||Aug 21, 2003||Jul 8, 2004||Dirk Noetzold||Method and system for valuation of complex systems, in particular for corporate rating and valuation|
|US20040138927||Jan 14, 2003||Jul 15, 2004||Alexander Eydeland||Method for producing a superior insurance model for commodity event risk|
|US20040153362||Jan 23, 2004||Aug 5, 2004||Progressive Casualty Insurance Company||Monitoring system for determining and communicating a cost of insurance|
|US20040186753||Mar 21, 2003||Sep 23, 2004||David Kim||System and method for catastrophic risk assessment|
|US20040199410||Jan 7, 2004||Oct 7, 2004||Hans Feyen||Method for evaluating flood plain risks|
|US20040220784||Apr 29, 2003||Nov 4, 2004||Stephenson David Mark||Automated generator of optimal models for the statistical analysis of data|
|US20040220837||Apr 30, 2003||Nov 4, 2004||Ge Financial Assurance Holdings, Inc.||System and process for a fusion classification for insurance underwriting suitable for use by an automated system|
|US20040220838||Apr 30, 2003||Nov 4, 2004||Ge Financial Assurance Holdings, Inc.||System and process for detecting outliers for insurance underwriting suitable for use by an automated system|
|US20040220839||Apr 30, 2003||Nov 4, 2004||Ge Financial Assurance Holdings, Inc.||System and process for dominance classification for insurance underwriting suitable for use by an automated system|
|US20040220840||Apr 30, 2003||Nov 4, 2004||Ge Financial Assurance Holdings, Inc.||System and process for multivariate adaptive regression splines classification for insurance underwriting suitable for use by an automated system|
|US20040230459||May 14, 2003||Nov 18, 2004||Dordick Rowan L.||Insurance for service level agreements in e-utilities and other e-service environments|
|US20040236611||Apr 30, 2003||Nov 25, 2004||Ge Financial Assurance Holdings, Inc.||System and process for a neural network classification for insurance underwriting suitable for use by an automated system|
|US20040236676||Mar 12, 2004||Nov 25, 2004||Kabushiki Kaisha Toshiba||Disaster risk assessment system, disaster risk assessment support method, disaster risk assessment service providing system, disaster risk assessment method, and disaster risk assessment service providing method|
|US20040249557||May 27, 2004||Dec 9, 2004||Wherenet Corp||Vehicle tag used for transmitting vehicle telemetry data|
|US20040249679||Oct 28, 2003||Dec 9, 2004||Risk Assessment Solutions, Llc||Systems and methods for qualifying expected loss due to contingent destructive human activities|
|US20050055248||Sep 4, 2003||Mar 10, 2005||Jonathon Helitzer||System for the acquisition of technology risk mitigation information associated with insurance|
|US20050055249||Sep 4, 2003||Mar 10, 2005||Jonathon Helitzer||System for reducing the risk associated with an insured building structure through the incorporation of selected technologies|
|US20050060141||Dec 9, 2003||Mar 17, 2005||Masaharu Suzuki||Information processing apparatus, information processing method and information processing program|
|US20050060207||Jul 7, 2004||Mar 17, 2005||Weidner James L.||Claims paid insurance|
|US20050070299||Sep 26, 2003||Mar 31, 2005||Siemens Information And Communication Networks, In||System and method for alternative presence reporting system|
|US20050091080||Oct 27, 2003||Apr 28, 2005||Biats Carl G.Jr.||System and method for managing liability insurer healthcare claims|
|US20050091085||Dec 6, 2004||Apr 28, 2005||Colley John L.||Method for evaluating the value of group and individual insurance products|
|US20050108063||Nov 5, 2003||May 19, 2005||Madill Robert P.Jr.||Systems and methods for assessing the potential for fraud in business transactions|
|US20050108066||Oct 6, 2004||May 19, 2005||Cooperation Of American Physicians||Property/casualty insurance and techniques|
|US20050125213 *||Dec 4, 2003||Jun 9, 2005||Yin Chen||Apparatus, system, and method for modeling and analyzing a plurality of computing workloads|
|US20050125259||Dec 5, 2003||Jun 9, 2005||Suresh Annappindi||Unemployment risk score and private insurance for employees|
|US20050131742||Nov 5, 2004||Jun 16, 2005||Hoffman Nancy J.||System and method for evaluating underwriting requirements|
|US20050137912||Mar 30, 2004||Jun 23, 2005||Rao R. B.||Systems and methods for automated classification of health insurance claims to predict claim outcome|
|US20050171885||Nov 3, 2004||Aug 4, 2005||Christman David T.||Object oriented demographically predictive insurance agency asset evaluation system and method|
|US20050192850||Mar 1, 2004||Sep 1, 2005||Lorenz Scott K.||Systems and methods for using data structure language in web services|
|US20050203828 *||Mar 12, 2004||Sep 15, 2005||Intelliclaim, Inc.||Insurance claim information system|
|US20050222867||Mar 31, 2004||Oct 6, 2005||Aetna, Inc.||System and method for administering health care cost reduction|
|US20050228692||Apr 8, 2004||Oct 13, 2005||Hodgdon Darren W||Incentive based health care insurance program|
|US20050234742||Feb 10, 2005||Oct 20, 2005||Hodgdon Darren W||Incentive based health care insurance program|
|US20050240451||Jun 13, 2005||Oct 27, 2005||Caelum Ip, Llc||System and method for generating a customized proposal in the development of insurance plans|
|US20050262039 *||May 20, 2004||Nov 24, 2005||International Business Machines Corporation||Method and system for analyzing unstructured text in data warehouse|
|US20050276401||Feb 28, 2005||Dec 15, 2005||Madill Robert P Jr||Systems and methods for assessing the potential for fraud in business transactions|
|US20060009289||Jul 6, 2005||Jan 12, 2006||Nintendo Co. Ltd.||Car-based entertainment system with video gaming|
|US20060015253||Oct 21, 2004||Jan 19, 2006||Yanhong Ochs||Risk management on the application of crop inputs|
|US20060015360||Oct 21, 2004||Jan 19, 2006||Yanhong Ochs||Insurance product associated with risk management on the application of crop inputs|
|US20060015373||Sep 10, 2003||Jan 19, 2006||Swiss Reinsurance Company||System and method for automated establishment of experience ratings and/or risk reserves|
|US20060015374||Oct 21, 2004||Jan 19, 2006||Yanhong Ochs||Risk management on the application of crop inputs|
|US20060053038||Sep 8, 2004||Mar 9, 2006||Warren Gregory S||Calculation of driver score based on vehicle operation|
|US20060064332||Nov 7, 2005||Mar 23, 2006||Michael Schoenbaum||Health cost calculator/flexible spending account calculator|
|US20060136273||Sep 9, 2005||Jun 22, 2006||Frank Zizzamia||Method and system for estimating insurance loss reserves and confidence intervals using insurance policy and claim level detail predictive modeling|
|US20060187889||Feb 19, 2005||Aug 24, 2006||Mr. Chand Mehta||System to mine information from data generated by Location Enabled Devices|
|US20060242046||Jun 30, 2005||Oct 26, 2006||American Express Travel Related Services Company, Inc.||Method and apparatus for consumer interaction based on spend capacity|
|US20060259333||May 16, 2005||Nov 16, 2006||Inventum Corporation||Predictive exposure modeling system and method|
|US20060287892||Jun 15, 2005||Dec 21, 2006||Hsb Solomon Associates, Llc||Insurance product, rating system and method|
|US20070016500||Mar 31, 2006||Jan 18, 2007||American Express Travel Related Services Co., Inc. A New York Corporation||Using commercial share of wallet to determine insurance risk|
|US20070016508||Jul 15, 2005||Jan 18, 2007||Pierre Lapointe||Method and system for redeemable vouchers using an insurance model|
|US20070016542 *||Jul 1, 2006||Jan 18, 2007||Matt Rosauer||Risk modeling system|
|US20070021987||Jul 21, 2006||Jan 25, 2007||Trurisk, Llc||Computerized medical modeling of group life insurance using medical claims data|
|US20070027725||Jul 29, 2005||Feb 1, 2007||Erwin Dirnberger||Insurance claim management|
|US20070027726||Apr 21, 2006||Feb 1, 2007||Warren Gregory S||Calculation of driver score based on vehicle operation for forward looking insurance premiums|
|US20070043656||Dec 30, 2005||Feb 22, 2007||Lancaster Clifton J||Method of risk modeling by estimating frequencies of loss and loss distributions for individual risks in a portfolio|
|US20070043662||Aug 24, 2006||Feb 22, 2007||The Hartford Steam Boiler Inspection And Insurance||Method of determining prior net benefit of obtaining additional risk data for insurance purposes via survey or other procedure|
|US20070100669||Nov 1, 2005||May 3, 2007||Accenture Global Services Gmbh||Collaborative intelligent task processor for insurance claims|
|US20070106539||Nov 27, 2006||May 10, 2007||Chris Gay||System and method for the assessment, pricing, and provisioning of distance-based vehicle insurance|
|US20070118399 *||Nov 22, 2005||May 24, 2007||Avinash Gopal B||System and method for integrated learning and understanding of healthcare informatics|
|US20080040676 *||Aug 10, 2006||Feb 14, 2008||Kabushiki Kaisha Toshiba||System and method for generating a customized workflow user interface|
|US20080126138 *||Nov 24, 2006||May 29, 2008||Eagle Insurance Agency, Inc.||System and method for presenting insurance offers|
|US20080236177 *||Feb 24, 2005||Oct 2, 2008||Danikin Inudstries Ltd.||Monitoring and Control System For Air Conditioner|
|1||Antonio et al. North American Actuarial Journal. 10:1, 30-48. Lognormal Mixed Models for Reported Claims Reserves. (Jan. 2006).|
|2||Apte et al. A probabilistic estimation framework for predictive modeling analytics. IBM Systems Journal. 41:3, 438-48. (2002).|
|3||Apte et al. Data-intensive analytics for predictive modeling. IBM Journal of Research and Development. 47:1, 17-23 (Jan. 2003).|
|4||Axelrod et al. Predictive Modeling in Health Plans. Abstract from Disease Management & Health Outcomes, 11:779-87(9). (Nov. 2003).|
|5||Bolstad, W. M. Introduction to Bayesian Statistics. A. John Wiley & Sons Inc. 2nd ed. 2007.|
|6||CAS Data Management and Information Educational Materials Working Party. Survey of Data Management and Data Quality Texts. Casualty Actuarial Society Forum, pp. 273-306. (Winter 2007).|
|7||Conger et al. Emphasis Apr. 2006. Predictive Modeling in Workers Compensation. pp. 18-21.|
|8||D'Arcy, Stephen P. Paper presented at World Risk and Insurance Economics Congress. Predictive Modeling in Automobile Insurance: A Preliminary Analysis. (2005).|
|9||de Alba, Enrique. Bayesian Estimation of Outstanding Claims Reserves. No. Ameri. Actuarial J. 6:4, 1-20. (2002).|
|10||de Ville, Barry. Decision Trees for Business Intelligence and Data Mining Using SAS Enterprise Miner. SAS Press Series. 2006.|
|11||Deloitte & Touche. Presentation Predictive Modeling for Claims: The New Claim Frontier (2006).|
|12||Derrig et al. Comparison of Methods and Software for Modeling Nonlinear Dependencies: A Fraud Application. (2006).|
|13||Ellingsworth et al. DM Review. Text Mining Improves Business Intelligence and Predictive Modeling in Insurance. (Jul. 2003).|
|14||Ellis et al. Applying Diagnosis-Based Predictive Models to Group Underwriting. Society of Actuaries, Issue 46, 1-7. (Aug. 2003).|
|15||Fellingham et al. Comparing Credibility Estimates of Health Insurance Claims Costs. No. Ameri. Actuarial J. 9:1, 1-12. (2005).|
|16||Fetterolf, Don. Paradise Lost: Return on Investment in Disease Management. Health Watch. Issue 52, pp. 14-17. (May 2006).|
|17||Grimes, Seth. The Word on Text Mining. Presentation. Portals, Collaboration, and Content Management. (Apr. 14, 2005).|
|18||Guszcza et al. Predictive Modeling for Property-Casualty Insurance. Presentation to SoCal Actuarial Club. (Sep. 22, 2004).|
|19||Guven, Serhat. Predictive Modeling. Future Fellows. (Jun. 2006).|
|20||Hong, S.J. et al. IBM Research Report RC-21570. Advances in Predictive Model Generation for Data Mining. (1999).|
|21||Macleod et al. Paper. Entropy-Reducing Properties of Predictive Financial Models. Aug. 27, 1992. Actuarial Research Clearing House. vol. 3 (1993).|
|22||Magnify Press Release. Erie Insurance Reduces Fraud Losses with FraudFocus(TM). Predictive Modeling Demonstrates Effectiveness for Auto, Property and Worker's Comp. (Feb. 4, 2005).|
|23||Magnify Press Release. Erie Insurance Reduces Fraud Losses with FraudFocus™. Predictive Modeling Demonstrates Effectiveness for Auto, Property and Worker's Comp. (Feb. 4, 2005).|
|24||Magnify Press Release. Magnify Applies Predictive Modeling to Worker's Comp Underwriting and Fraud Detection. Chicago, IL (Mar. 1, 2005).|
|25||Meyers, Glenn. On Predictive Modeling for Claim Severity. Casualty Actuarial Society Forum. pp. 215-253. (Spring 2005).|
|26||Monsour, Christopher. Presentation: Discrete Predictive Modeling. Casualty Actuarial Society. Special Interest Seminar on Predictive Modeling. (Oct. 5, 2004).|
|27||Monsour, Christopher. Presentation: What is Predictive Modeling? (2005).|
|28||Morgan et al. Conjugate Bayesian Analysis of the Negative Binomial Distribution. Actuarial Research Clearing House. vol. 1, pp. 97-118, (1993).|
|29||Mosley, R. The Use of Predictive Modeling in the Insurance Industry. Pinnacle Actuarial Resources Inc. (Jan. 2005).|
|30||Pednault et al. IBM Research Report RC-21731. Handling Imbalanced Data Sets in Insurance Risk Modeling. (Mar. 10, 2000).|
|31||Pednault et al. IBM Research Report RC-21757. The Importance of Estimation Errors in Cost-Sensitive Learning. (May 30, 2000).|
|32||Predictive Modeling Applications. Weyuker, L. & Minnich, J. Record, 31:2. New Orleans Health/Pension Spring Meeting, Session 3PD. (Jun. 2005).|
|33||Predictive Modeling Raises Opportunities and Issues for Actuaries and Insurers, CAS Annual Meeting is Told. Insurance Newscast (Dec. 15, 2005).|
|34||Predictive Modeling. Record, 28:2. Spring Meeting, San Francisco, CA. Session 99OF. (Jun. 2002).|
|35||Predictive Modeling-Current Practices and Future Applications. Record, 30:1. Spring Meeting, Anaheim, CA. Session 64PD. (May 2004).|
|36||Roberts, Jack. Risk & Insurance. Fighting CAT risk creep. 18:8, p. 26(2). (2007).|
|37||Rosenberg et al. Predictive Modeling with Longitudinal Data: A Case Study of Wisconsin Nursing Homes. (Feb. 4, 2006).|
|38||Roudebush et al. Converting Clinical Literature to an Insured Population: A Comparison of Models Using Nhanes. No. Ameri. Actuarial J. 6:4, 55-66 (2002).|
|39||Sanche et al. Variable Reduction for Predictive Modeling with Clustering. Casualty Actuarial Society Forum, pp. 89-100. (Winter 2006).|
|40||Sharp, Keith P. Aspects of Interest Rate Models. Actuarial Research Clearing House. vol. 1, pp. 433-457. (Aug. 25, 1990).|
|41||Steed, Judy. Winning Ways. Toronto Star, p. 3-4 (May 21, 2007).|
|42||Stehno, Chris E. What We Have Learned in the Last 50 Years-and Aren't Using. Health Watch. Issue 52, pp. 18-21. (May 2006).|
|43||Table of Contents of White Paper. Predictive Modeling in Insurance: An insurance industry executive briefing. SAS (Predictive Modeling in Insurance), publisher. (Mar. 23, 2007).|
|44||Wang, Wei. Predictive Modeling Based on Classification and Pattern Matching Methods. Thesis. (May 1999).|
|45||*||Weizhong Yan, "Designing a Neural Network Decision System for Automated Insurance Underwriting," Int'l Joint Conference on Neural Networks, Jul. 2006, pp. 1-8.|
|46||Werner et al. GLM Basic Modeling: Avoiding Common Pitfalls. Casualty Actuarial Society Forum. pp. 257-272. (Winter 2007).|
|47||Wolak, Dan. An Actuarial Response to the Health-Care Crisis. Society of Actuaries. Issue 47, 1-9. (Apr. 2004).|
|48||Woodfield, Terry J. Paper 071-30. Predicting Workers' Compensation Insurance Fraud Using SASŪ Enterprise Miner(TM) 5.1 and SASŪ Text Miner. 2004.|
|49||Woodfield, Terry J. Paper 071-30. Predicting Workers' Compensation Insurance Fraud Using SASŪ Enterprise Miner™ 5.1 and SASŪ Text Miner. 2004.|
|50||Woodley et al. Assessing Predictive Modeling Tools for Pricing and Underwriting. Health Watch. Issue 51, pp. 30-33. (Jan. 2006).|
|51||Wu et al. Paper. Casualty Actuarial Society Forum. pp. 113-138. Does Credit Score Really Explain Insurance Losses? Multivariate Analysis from a Data Mining Point of View. 2003.|
|52||Wu, Cheng-sheng Peter et al. A View Inside the "Black Box:" A Review and Analysis of Personal Lines Insurance Credit Scoring Models Filed in the State of Virginia. Casualty Actuarial Society Forum, pp. 251-290 (Winter 2004).|
|53||Young, Virginia R. Actuarial Research Clearing House. vol. 1. Robust Bayesian Credibility Using Semiparametric Models. (1999).|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8112377 *||May 15, 2008||Feb 7, 2012||Sap Ag||Client-side rule engine for executing business rules in rich internet applications|
|US8355934 *||Jan 25, 2010||Jan 15, 2013||Hartford Fire Insurance Company||Systems and methods for prospecting business insurance customers|
|US8473454 *||Mar 10, 2009||Jun 25, 2013||Xerox Corporation||System and method of on-demand document processing|
|US8543445 *||Dec 21, 2009||Sep 24, 2013||Hartford Fire Insurance Company||System and method for direct mailing insurance solicitations utilizing hierarchical bayesian inference for prospect selection|
|US8655682||Jul 15, 2011||Feb 18, 2014||Gitika Srivastava||Treatment decision engine with applicability measure|
|US8694340 *||Feb 21, 2012||Apr 8, 2014||United Services Automobile Association||Systems and methods to evaluate application data|
|US8706521||Jul 15, 2011||Apr 22, 2014||Naresh Ramarajan||Treatment related quantitative decision engine|
|US8892452 *||Nov 9, 2012||Nov 18, 2014||Hartford Fire Insurance Company||Systems and methods for adjusting insurance workflow|
|US8996359||May 2, 2012||Mar 31, 2015||Dw Associates, Llc||Taxonomy and application of language analysis and processing|
|US9020807||Jan 17, 2013||Apr 28, 2015||Dw Associates, Llc||Format for displaying text analytics results|
|US20100235305 *||Sep 16, 2010||Xerox Corporation||System and method of on-demand document processing|
|US20110153419 *||Jun 23, 2011||Hall Iii Arlest Bryon||System and method for intelligent modeling for insurance marketing|
|US20110184766 *||Jan 25, 2010||Jul 28, 2011||Hartford Fire Insurance Company||Systems and methods for prospecting and rounding business insurance customers|
|US20120117019 *||May 10, 2012||Dw Associates, Llc||Relationship analysis engine|
|US20120323609 *||Dec 20, 2012||Enservio, Inc.||Systems and methods for predicting the value of personal property|
|US20140122134 *||Oct 22, 2013||May 1, 2014||Enservio, Inc.||Systems and methods for predicting the value of personal property|
|US20140188522 *||Dec 19, 2013||Jul 3, 2014||Enservio, Inc.||Systems and methods for predicting the value of personal property|
|US20150206249 *||Dec 5, 2014||Jul 23, 2015||Enservio, Inc.||Systems and methods for predicting the value of personal property|
|WO2013009599A2 *||Jul 6, 2012||Jan 17, 2013||The Travelers Indemnity Company||Systems and methods for business classification|
|WO2013009599A3 *||Jul 6, 2012||Mar 7, 2013||The Travelers Indemnity Company||Systems and methods for business classification|
|U.S. Classification||705/35, 705/4|
|Cooperative Classification||G06Q40/00, G06Q40/08|
|European Classification||G06Q40/08, G06Q40/00|
|Dec 20, 2007||AS||Assignment|
Owner name: HARTFORD FIRE INSURANCE COMPANY, CONNECTICUT
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:KENEFICK, TIMOTHY PAUL;BESMAN, ERIC;MURPHY, THERESA COTE;REEL/FRAME:020277/0629
Effective date: 20071220
|Aug 27, 2014||FPAY||Fee payment|
Year of fee payment: 4